I just installed KDE 4.3 and it's looking good. Some features returned that I was missing. You can now once again display applications by name rather than description in the K-menu. You can now enable a nice kcontrol-like tree-view in the System Settings.

Some things are still missing though... like different wallpapers on different desktops. There are "Activities" which can have different wallpapers, but I can't for the life of me figure out how I'm supposed to be using them. I also lag and/or crash every time I Zoom Out in the cashew, possibly thanks to 3840x1200 screen resolution. I'm going to assume Activities are still a work-in-progress.

I was horrified to open Kopete and see that configuring the contact list window now uses the same completely broken configuration dialog that Amarok 2 uses for their playlist. Oh how I hope someone rethinks this.

There's a new Qt and Plasma theme in KDE 4.3 that looks pretty nice. Overall every release of KDE4 seems to become more stable, more polished, more eye-candy (if you want it).

Well then. I’ve been motivated by Hans Chen who originally decided to walk the path to a KDE developer and to do my own. For the technophobes, KDE is an actively developed desktop interface (for want of a better description) which is pushing ahead what the desktop is capable. It is also a community for everybody – the programmers, the artists, the PR folks and most importantly, the users.

Somebody once said that open-source will truly succeed when anybody and everybody will have the capability to create the environment around them completely as they see fit. I’m not denying that this’ll result in a lot of crappy environments (or not to my liking) but it basically says contribution shouldn’t be limited to those who mutter binary in their sleep.

I’m going to see if I am able to make this contribution. Let’s start by giving you the case study:

I do know PHP and do web-development. I have coded a black jack CLI game in C++ (well it was a start). I do graphics design and am proficient in The GIMP. I once touched Python but probably have forgotten a lot. I am a student and have no intention of going into programming as a career and have been self-taught. I have no deadlines, no clear goals nor any roadmap for this project, and am always juggling a variety of other projects on the side.

That, sounding quite like the commitment level proffered by most individuals seems like a vague and – well, truly realistic to be honest. I shall post my progress here with no fixed schedule and see how things go along

I’m thinking of starting by reading “Acclerated C++” – mainly because whoever thought up that title got the 100% keyword efficiency award for word estate. 2 buzzwords in 2 words. Eeeexceeelent.

No related posts.

Best Skins Ever is a great company that provides thin “skins” for all kinds of electronic devices. From cameras to portable gaming devices to cell phones to laptops, Best Skins Ever covers them all.

After my old iPhone case broke and I purchased a brand new iPhone 3GS, I decided that I needed to replace the broken case. I paid $30 for that case at the Apple store and it lasted about 6 months. I was extremely disappointed with its quality. The rubber around the headphone jack was what finally broke, but the rubber around the bottom bezel of the phone was also wearing thin. Overall, it was a bad experience.

And then I ran into Best Skins Ever. I read reviews for their skins and saw not even one negative review of their skins. I read that their skins were reasonably priced, their durability was great, and their customer service was exceptional. I even heard that people who botched their install received free replacements.

These skins are unlike ordinary skins. They contain an adhesive that sticks to your phone permanently, but does not damage the phone when removed. Any attempt to move the skin once it has dried will stretch the skin. Aside from the application process, these skins are perfect.

So I decided to obtain two of them. They had many different options on their iPhone page (which I don’t describe here); and I chose the hardest one to apply (but also the best one once applied). The skins are $7.99 a piece. Much cheaper than $30 for that bulky case that I bought at the Apple store! So I bought two just in case I messed up the first one.

The install process requires that you soak the skin with soapy water, so that you can adjust it as you place it on your device and so that you do not get fingerprints on the sticky side of the skin. Once the skin dries onto the phone, it sticks firmly; all dust and bubbles are permanently trapped in there. I went through the install, not really knowing what to do. Amazingly, I was able to install it about 90% perfectly on my first try. It took me about 5 minutes to put the front screen protector on (it even covers the home button!) and about an hour to situate the back skin. The reason it takes so long is that it wraps around the bezel, around the top SIM slot, and around the bottom charge port.

I ended up with some small bubbles and a bit of blue fuzz stuck behind the back of the phone, but that is no big deal; especially when compared to having no protection at all and ending up with tons of scratches. The screen protector is not as smooth as the plastic screen protectors and it causes a little extra glare at some angles. But the protection that the back receives is amazing. My phone is now thinner than it ever was with a case and it is much more comfortable to hold. It’s great that I can now see the phone in an almost case-less state.

I highly recommend purchasing a BSE for any device that you want minimal protection!

Pictures (although not the best) are located below. As you can see, my install is not perfect. I have some small bubbles (those are not scratches) on the bottom bezel. Also, some of the outer wrappings overlap.

Ok, hopefully this will be the last mention of this for a bit. I've gone back and completely rewritten and expanded greatly on my old Git tutorial. It's still at the same place, Mindstab's Git Guide, just with a new name. It still needs lots more work and expanding, which will eventually happen, but probably not right now. As it is it's now good enough to let stand on its own while I get back to cl-pack. It has everything you need to set up a repository and get it on a server in one of many configurations. It just doesn't have a lot on longer term development and multi team management of multiple repositories and branches. Haha, I also recast it as an Alice and Bob story.

Recently I worked on some new project, and as always I created a local Git repository as a start. After working on it several days, creating lots of commits, I had to publish it into the central Subversion repository (which is one of the VCSs we got). I could have done this by creating a new folder in SVN and add the latest version of all files of the project to it, but that way all history would be gone, which I didn’t like.

Git has a feature to work with SVN repositories, git-svn, but that’s intended to check out existing code from SVN and work on it, not publishing an existing Git tree into a Subversion repository.

A first rather naive approach didn’t work out (as somewhat expected), but then I figured out how to achieve this anyway.

As a test, let’s first create an empty SVN repository and a Git repository with some commits:

$ svnadmin create repo
$ svn co file:///Users/nicolas/Temp/git_to_svn/repo svn_repo
Checked out revision 0.
$ cd svn_repo
$ svn mkdir trunk tags branches
A         trunk
A         tags
A         branches
$ svn commit -m "Create repository structure"
Adding         branches
Adding         tags
Adding         trunk

Committed revision 1.
$ cd ..

$ mkdir project; cd project
$ git init
Initialized empty Git repository in /Users/nicolas/Temp/git_to_svn/project/.git/
$ echo "foo" > test.txt; git add test.txt; git commit -m "Initial version"
master (root-commit) 88464cf] Initial version
 1 files changed, 1 insertions(+), 0 deletions(-)
 create mode 100644 test.txt
$ echo "bar" > test.txt; git commit test.txt -m "Second version"
master cb62866] Second version
 1 files changed, 1 insertions(+), 1 deletions(-)

We now can set up git-svn:

$ git svn init -s file:///Users/nicolas/Temp/git_to_svn/repo/
$ git svn fetch
r1 = 741ab63aea786882eafd38dc74369e651f554c9c (trunk)

Depending on the layout of your SVN project, you might need to drop the -s parameter and add -t, -T or -b flags, see the git-svn manpage.

A little naive we could try to push everything to the SVN repository now:

$ git svn dcommit
Unable to determine upstream SVN information from HEAD history.
Perhaps the repository is empty. at /opt/local/libexec/git-core/git-svn line 439.

This fails since the git svn command can’t figure out which commits to push: there’s no link between our original Git repository and the Subversion heads.

To fix this, we can use a Git graft to link them. We’ll tell Git the commit which created the SVN folder in which we want to store the project is the parent commit of the first commit in our Git repository:

$ git show-ref trunk
741ab63aea786882eafd38dc74369e651f554c9c refs/remotes/trunk
$ git log --pretty=oneline master | tail -n1
88464cfdf549a82b30ee7c52e53e2b310f0d9ec4 Initial version
$ echo "88464cfdf549a82b30ee7c52e53e2b310f0d9ec4 741ab63aea786882eafd38dc74369e651f554c9c" >> .git/info/grafts

If now you execute git log, you’ll see the “Create repository structure” SVN commit is displayed after our “Initial version” commit.

Pushing to SVN now works fine:

$ git svn dcommit
Committing to file:///Users/nicolas/Temp/git_to_svn/repo/trunk ...
	A	test.txt
Committed r2
	A	test.txt
r2 = 8c72757dd3a7d550ed8ef393bb74c0350d22dbac (trunk)
No changes between current HEAD and refs/remotes/trunk
Resetting to the latest refs/remotes/trunk
test.txt: locally modified
	M	test.txt
Committed r3
	M	test.txt
r3 = ca0fc06d477bcd4dd5c6f6d2ae6d94356b510280 (trunk)
No changes between current HEAD and refs/remotes/trunk
Resetting to the latest refs/remotes/trunk

All set

Well folks, here’s another quick tech tip that I use once in a while. How do you rip only the audio from an .FLV file? .FLV files, or Flash Video files are the format used in browser-embedded videos, common on video-sharing sites such as YouTube or Vimeo (and Eadrax!) For whatever reason if you have an .FLV file of your favourite music video, now you can get the music rocking solo.

mencoder a.flv -o a.mp3 -of rawaudio -oac mp3lame -lameopts cbr:br=192 -ovc copy

I’ll stay off tech tips for a while as the weekend is coming up.

Related posts:

1. Tech tip #2: MPlayer play music recursively in a directory.
2. Tech Tip: Speed up Firefox to use less CPU.

I have always wondered how to do this. It’s quite often I have directories full of media files (specifically music) and subdirectories within them also with music files, and though the manual for MPlayer is thicker than the Chinese phonebook I have not been able to find any option for it.

I did however find a method which isn’t exactly the guru one-liner, but here it is anyway. It’s broken into two steps, the first to create a playlist:

find -maxdepth 1 -type f -name \*.\* > playlist

Then finally play the playlist:

mplayer -playlist playlist

Just add a -loop 0 suffix if you want to loop

Related posts:

1. Tech tip #3: Rip audio from an .FLV file.
2. Music Composition: Surprises
3. Music Composition: The Spice of Life

For about a year now I have had a very annoying problem with my file server. One of the two Seagate drives simply disappeared from time to time. It was never the same one and I couldn't tell what pattern the disappearances followed. The system log looked something like:

Jul 11 18:55:02 hostname ata5.00: exception Emask 0x10 SAct 0x1 SErr 0x190002 action 0xe frozen
Jul 11 18:55:02 hostname ata5.00: edma_err_cause=00000020 pp_flags=00000003, SError=00180000
Jul 11 18:55:02 hostname ata5: SError: { RecovComm PHYRdyChg 10B8B Dispar }
Jul 11 18:55:02 hostname ata5.00: cmd 61/80:00:a9:47:55/00:00:1b:00:00/40 tag 0 ncq 65536 out
Jul 11 18:55:02 hostname res 40/00:00:a9:47:55/00:00:1b:00:00/40 Emask 0x10 (ATA bus error

At first I assumed that it was due to a faulty SATA controller on my motherboard. I found a PCI-e card that I could fit into my motherboard. It seemed to work just fine for a while, apart from the hassle with motherboard not being able to boot directly from it and being forced to use a noisy old IDE drive for storing grub and kernel to boot from.
After a month or so, one of the hard drives popped out again. Dang. This time I started trying to diagnose my Seagate discs. They have excellent diagnostic tools that runs under linux on their website. I even sent two of them back to seagate and got them refurbished.
After a couple of months or so again it was time to resync the RAID due to discs popping out. Man, I was frustrated with this by now. Since the only things left was cabling and PSU by now and I found this. I decided to go for a new PSU, I ripped out the old FSP and put in a new Corsair. Corsair really has become my favourite ones over time.
According to the info I found this problem should reappear during load. Since I've done a resync of the raid, have a couple of full backup jobs run and a bunch of my pictures uploaded to it, it really seems rock stable now =).

Looks like Microsoft releases the Linux drivers to enable a Linux kernel running as a guest in a Hyper-V hypervisor to run in ‘enlightened mode’, which sounds pretty much like Xen’s PV drivers for Windows, providing better IO performance, under the GPLv2 (which is the same open-source license as the Linux kernel itself). Quoting the Hyper-V Architecture and Feature Overview:

Enlightened I/O is a specialized virtualization-aware implementation of high level communication protocols (such as SCSI) that utilize the VMBus directly, bypassing any device emulation layer. This makes the communication more efficient but requires an enlightened guest that is hypervisor and VMBus aware.

The drivers seem to be developed by Novell, so I guess the Boycott Novell guys will have some more coverage^Wrants soon (Update: can’t find the reference on this anymore, so this might be a false statement, sorry. Thanks for pointing out RubenV)

Interesting times on the virtualization front… Although I for one do not plan to replace Xen, xVM or VirtualBox anytime soon.

Sources:

• Microsoft Press Release
• Blog of the 451 Group
• LKML

On a side note: Red Hat entered the Standard & Poor’s 500 index, which might show Linux is gaining more interest from enterprises and investors.

• We're using the pure JavaScript RSA and BigInteger library from ohdave.com/rsa. I make no claims to the speed of efficiency, and I have no real benchmarks of just this library, either. We searched "rsa javascript" and went with the first link we found on Google. We have not investigated optimizing this, though we probably should.
  
• The actual requests and responses are pretty small - no more than 300 bytes after the initial data loading.
• These requests occur every 1.5 seconds, but we'll likely increase that down the road.
• The testing was done in a VM running a 64-bit copy Windows Vista. The VM host is an Intel i7 920 running qemu-kvm (or, it is a really fast VM, thanks to CPU virtualization extensions.).
• Browsers: Firefox 3.5.1 with the JIT compiler enabled; Google Chrome 2.0.172.37; Safari 4.0.2; all under Windows Vista - all being the latest version available. (I wish I could have tested Opera too, but it has a very very strange bug involving the RSA.)
• I'm going to say "99% CPU," "10% CPU," and so on. Task Manager splits the load percentage over every core on the machine, and the i7 has 8 that it detects. However, as stated, this was done in a VM, which I allocated a single CPU to. "99% CPU" is going to mean "hey this app just froze on me and took most of the OS with it."
  

• Firefox jumps from 99% CPU 'down' to 65% CPU as reported by Vista's task manager. On occasion it hit 0, but only for very brief periods of time. The RAM usage was also fun to watch: over a period of many requests, it would mostly remain stable - only to jump by 80MB or more for several seconds, and then back down. This resulted in every three out of four keystrokes being lost, on average, never mind the huge memory churn.
  
• Safari sat there and looked at us funny.

• Chrome shocked my friend and I. Sure, we had heard that it was fast, and I had done a bit of reading on it. However, I was quite surprised when I saw that it was using a maximum of 10% CPU and roughly 18MB of RAM without any spikes in either. Wow.

I've got a bunch of stuff that I need to get rid of, both because it's taking up precious space, and because I'd rather have the cash right now. I've got the stuff listed on a few forums, but it's a tough crowd. I figured some friends or Linux folk might be interested in some of the hardware.

• Palit Radeon� HD 4870 Sonic Dual Edition (512MB) - Product Page
• Highpoint ROCKETRAID 2640X4 - Product Page
• HIGHPOINT ROCKETRAID 3510 - Product Page

Everything's Linux compatible, and I'm not looking for much for any these items ($100-200 a pop). If you have any interest, leave a comment below with a valid e-mail address. And don't worry, I'm the only one who can see the addresses.

Honestly, KDE 4 is the best desktop I've used to date. It's still rough around a few edges, but perfectly usable. I've said all this before. Not much bothers me about the roughness except a few issues in Plasma that make it a pain at times.

After KDE has been running for a few days, Plasma will randomly go up to 100% CPU usage on one of my cores. Well, I have three other CPU cores to use, but while at 100% Plasma is unusable. Slightly annoying, no? I'm not sure what the issue is. A Google search reveals that other people are having similar issues, but there aren't many fixes if any.

I'm currently using 4.2.2, which isn't the newest in Gentoo's repository, but I don't see the necessity to upgrade to 4.2.4 when the 4.3 release is just around the corner by the end of this month. I'm hoping that fixes this issue.

Although I'm no KDE developer, I wonder whether Plasma is threaded or not? Might make a difference in this case...

Soooo, I lost my job back on May 20th. It's been nearly two months, and I really haven't been able to find a new one just yet. I've focused most of my time applying for restaurant management positions, and I did get a ton of interviews. I also didn't get a single job offer. Don't get me wrong, I know a great deal about running restaurants and bars. Might only be so young, but the experience and knowledge is solid. Call my references and they'll say the same thing.

The downside of getting a management position at this point of the year would mean canceling my vacation with three of my friends at the end of August. Most restaurants require you work there six months to a year before you get a week's vacation. So after the last few rejections I received, I decided that I'd go on my merry way and find an easy serving or bartending job that would hold me over until vacation was done and over with.

I'm alright on the money situation because I have a ton of crap I've been selling on eBay, and my dad said he'd take care of the student loan bills until I'm back on my feet. I'm not deliberately trying to stay off my feet, but we've had this vacation booked since February. It'd be a shame to ruin it now.

In the meantime, I'm having some business cards printed at VistaPrint. I'm going to pass these business cards out amongst friends, family, and bars I frequent. The cards are advertising cheap computer repairs and upgrades. While it won't be great business, I figure it's something to help pay the bills until I do find a suitable job.

On my Macbook (4,1) I am currently using Debian with kernel 2.6.30-1-686-bigmem. This Macbook has Broadcom 4328 wireless chipset installed (02:00.0 Network controller: Broadcom Corporation BCM4328 802.11a/b/g/n (rev 03)) and unfortunately the necessary kernel module provided by Broadcom is pretty unstable. Or very unstable. Oh well…it’s totally unstable.

I had random freezes, usually when I first booted and tried to modprobe the module. After some searching around the net and a lot of experiments I’ve managed to create a kernel module that looks quite stable. At least I stopped getting any more lockups and freezes…To reproduce the module with the patches I’ve used follow the directions bellow step by step.

Create necessary dirs:
mybox:~# mkdir hybrid_wl
mybox:~# cd hybrid_wl

Download drivers package from Broadcom:
802.11 Linux STA 32-bit Driver
mybox:~/hybrid_wl# wget http://www.broadcom.com/docs/linux_sta/hybrid-portsrc-x86_32-v5_10_91_9.tar.gz

Download a few more patches from Archlinux and Gentoo:
hidden-essid patch
2.6.30 patch 1
2.6.30 patch 2
hybrid-portsrc-x86_32-v5_10_91_9-convert_to_net_device_ops.diff
mybox:~/hybrid_wl# wget http://aur.archlinux.org/packages/broadcom-wl/broadcom-wl/hidden-essid.patch
mybox:~/hybrid_wl# wget http://aur.archlinux.org/packages/broadcom-wl/broadcom-wl/broadcom-sta-5.10.91.9-linux-2.6.30.patch
mybox:~/hybrid_wl# wget http://aur.archlinux.org/packages/broadcom-wl/broadcom-wl/broadcom-sta-5.10.91.9-linux-2.6.30-2.patch
mybox:~/hybrid_wl# wget -O hybrid-portsrc-x86_32-v5_10_91_9-convert_to_net_device_ops.diff http://bugs.gentoo.org/attachment.cgi?id=195182

Extract package:
mybox:~/hybrid_wl# tar -xzf /path/to/hybrid-portsrc-x86_32-v5_10_91_9.tar.gz

Start Patching:
mybox:~/hybrid_wl# sed -i hidden-essid.patch -e 's|5.10.79.10|src/wl/sys|g'
mybox:~/hybrid_wl# patch -p0 < hidden-essid.patch
patching file src/wl/sys/wl_iw.c
mybox:~/hybrid_wl# sed -i broadcom-sta-5.10.91.9-linux-2.6.30.patch -e 's|hybrid-portsrc-x86_32-v5_10_91_9.orig/||g'
mybox:~/hybrid_wl# sed -i broadcom-sta-5.10.91.9-linux-2.6.30.patch -e 's|hybrid-portsrc-x86_32-v5_10_91_9/||g'
mybox:~/hybrid_wl# patch -p0
patching file src/wl/sys/wl_iw.c
Hunk #1 succeeded at 611 (offset 1 line).
Hunk #2 succeeded at 640 (offset 1 line).
Hunk #3 succeeded at 1119 (offset 1 line).
Hunk #4 succeeded at 1147 (offset 1 line).
Hunk #5 succeeded at 1807 (offset 1 line).
Hunk #6 succeeded at 1942 (offset 1 line).
patching file src/wl/sys/wl_linux.c
patching file src/wl/sys/wl_linux.h
mybox:~/hybrid_wl# patch -p0
patching file src/wl/sys/wl_linux.c
mybox:~/hybrid_wl# sed -i hybrid-portsrc-x86_32-v5_10_91_9-convert_to_net_device_ops.diff -e 's|a/src/|src/|g'
mybox:~/hybrid_wl# sed -i hybrid-portsrc-x86_32-v5_10_91_9-convert_to_net_device_ops.diff -e 's|b/src/|src/|g'
mybox:~/hybrid_wl# patch -p0
patching file src/wl/sys/wl_linux.c
Hunk #1 succeeded at 225 (offset 6 lines).
patching file src/wl/sys/wl_iw.c

Compile the kernel module:
mybox:~/hybrid_wl# make -C /lib/modules/2.6.30-1-686-bigmem/build M=`pwd` clean
make: Entering directory `/usr/src/linux-headers-2.6.30-1-686-bigmem'
make: Leaving directory `/usr/src/linux-headers-2.6.30-1-686-bigmem'
mybox:~/hybrid_wl# make -C /lib/modules/2.6.30-1-686-bigmem/build M=`pwd`
make: Entering directory `/usr/src/linux-headers-2.6.30-1-686-bigmem'
LD /root/hybrid_wl/built-in.o
CC [M] /root/hybrid_wl/src/wl/sys/wl_linux.o
CC [M] /root/hybrid_wl/src/wl/sys/wl_iw.o
CC [M] /root/hybrid_wl/src/shared/linux_osl.o
LD [M] /root/hybrid_wl/wl.o
Building modules, stage 2.
MODPOST 1 modules
WARNING: modpost: missing MODULE_LICENSE() in /root/hybrid_wl/wl.o
see include/linux/module.h for more information
CC /root/hybrid_wl/wl.mod.o
LD [M] /root/hybrid_wl/wl.ko
make: Leaving directory `/usr/src/linux-headers-2.6.30-1-686-bigmem'

Install the new module:
mybox:~/hybrid_wl# cp wl.ko /lib/modules/2.6.30-1-686-bigmem/kernel/drivers/net/wireless/
mybox:~/hybrid_wl# depmod
mybox:~/hybrid_wl# modprobe wl

Check if everything loads correctly:
mybox:~/hybrid_wl# dmesg |tail
[ 66.229797] lib80211: common routines for IEEE802.11 drivers
[ 66.229805] lib80211_crypt: registered algorithm 'NULL'
[ 66.301793] wl: module license 'unspecified' taints kernel.
[ 66.301802] Disabling lock debugging due to kernel taint
[ 66.305919] wl 0000:02:00.0: PCI INT A -> GSI 16 (level, low) -> IRQ 16
[ 66.305933] wl 0000:02:00.0: setting latency timer to 64
[ 66.406146] lib80211_crypt: registered algorithm 'TKIP'
[ 66.408646] eth1: Broadcom BCM4328 802.11 Wireless Controller 5.10.91.9
[ 76.524135] eth1: no IPv6 routers present

You can also chek the iwconfig output. Hopefully everything will be fine…
I hope this saves a few hours of searching and experimenting for some people…

References:
1) 802.11 Linux STA driver
2) AUR broadcom-wl 5.10.91.9-2
3) Gentoo Bug: 284450 (New ebuild: net/wireless/broadcom-sta)

• Ryan’s Tech Blog » HOWTO: start using lisp in your work environment (part 1)
  (tags: lisp, work, howto)

(This is a reply on a post by Ahmed Soliman on recursion performance in (C)Python, and CPython function call overhead in general. I started to write this as a comment on his post, but it turned out much longer, so sending it over here in the end.)

Hey,

As discussed before, this is not a fair comparison, since the non-recursive version is much ’smarter’ than the recursive one: it calculates values and will never recalculates them, whilst the recursive version calculates everything over and over again.

Adding some simple memoization helps a lot. First, my testing code:

Here are the benchmarks on my MacBook Pro Intel Core2Duo 2.33GHz with 3GB RAM (running quite a lot of applications). Do note the ‘dumb’ version calculates fib(35), whilst the slightly optimized versions, which still use recursion but much less recursive calls (as they should) or your second version calculate fib(150).

Using MacOS X 10.5.6 stock CPython 2.5.1:

MacBook:Projects nicolas $ python -V
Python 2.5.1

MacBook:Projects nicolas $ python fib.py 35 150
fib(35) = 9227465
Calculation took 12.8542108536 seconds

Calculating the amount of recursive calls to calculate fib(35)
Calculating fib(35) = 9227465 took 29860703 calls

fib2(150) = 9969216677189303386214405760200
Calculation took 0.00020694732666 seconds

memoize_dict(fib)(150) = 9969216677189303386214405760200
Calculation took 0.00141310691833 seconds

memoize_constant_list(151, fib)(150) = 9969216677189303386214405760200
Calculation took 0.000310182571411 seconds

Overall it looks like fib2 and memoize_constant_list perform fairly similar, I guess function call overhead and list.append have a similar influence on performance in this case.

Using Jython 2.5.0 from the binary distribution on the Java HotSpot 64bit Server VM as shipped for OS X 10.5.6:

MacBook:Projects nicolas $ ./Jython/jython2.5.0/jython -V
Jython 2.5.0

MacBook:Projects nicolas $ ./Jython/jython2.5.0/jython fib.py 35 150
fib(35) = 9227465
Calculation took 12.5539999008 seconds

Calculating the amount of recursive calls to calculate fib(35)
Calculating fib(35) = 9227465 took 29860703 calls

fib2(150) = 9969216677189303386214405760200
Calculation took 0.0519998073578 seconds

memoize_dict(fib)(150) = 9969216677189303386214405760200
Calculation took 0.00399994850159 seconds

memoize_constant_list(151, fib)(150) = 9969216677189303386214405760200
Calculation took 0.00300002098083 seconds

The ‘dumb’ fib implementation performs similar in both CPython and Jython. Jython performs significantly less good on the other implementations though, but maybe todays news could help here, not sure how much locking on dict and list access Jython introduces.

Finally, using Unladen Swallow 2009Q2, self-compiled from SVN on the same system, using standard settings:

MacBook:Projects nicolas $ ./unladen-swallow/unladen-2009Q2-inst/bin/python -V
Python 2.6.1

MacBook:Projects nicolas $ ./unladen-swallow/unladen-2009Q2-inst/bin/python fib.py 35 150
fib(35) = 9227465
Calculation took 12.2675719261 seconds

Calculating the amount of recursive calls to calculate fib(35)
Calculating fib(35) = 9227465 took 29860703 calls

fib2(150) = 9969216677189303386214405760200
Calculation took 0.000118970870972 seconds

memoize_dict(fib)(150) = 9969216677189303386214405760200
Calculation took 0.000972986221313 seconds

memoize_constant_list(151, fib)(150) = 9969216677189303386214405760200
Calculation took 0.00036096572876 seconds

which is similar to, slighly better or slightly worse than the CPython run, and when enforcing JIT (which introduces a significant startup time, which is not measured here):

MacBook:Projects nicolas $ ./unladen-swallow/unladen-2009Q2-inst/bin/python -j always fib.py 35 150
fib(35) = 9227465
Calculation took 14.6129109859 seconds

Calculating the amount of recursive calls to calculate fib(35)
Calculating fib(35) = 9227465 took 29860703 calls

fib2(150) = 9969216677189303386214405760200
Calculation took 0.0432291030884 seconds

memoize_dict(fib)(150) = 9969216677189303386214405760200
Calculation took 0.0363459587097 seconds

memoize_constant_list(151, fib)(150) = 9969216677189303386214405760200
Calculation took 0.0335609912872 seconds

which, to my surprise, performs pretty worse than the default settings.

Overall: your first implementation performs tons and tons of function calls, whilst the second one, which resembles memoize_list_fib in my code (which is recursive), performs significantly less function calls and in the end memoize_list_fib performs almost as good as your second version (it performs +- the same number of function calls as the number of times you’re going through your loop).

So whilst I do agree function calls in Python are reasonably slow compared to plain C function calls (which is just a jmp, no frame handling etc. etc. required), your comparison between your recursive and non-recursive implementation is completely unfair, and even if calculating fib(35) takes several seconds, consider you’re doing a pretty impressive 29860703 function calls to perform the calculation.

Time to get some sleep.

Last night I added  getdelim(3) and getline(3) to  NetBSD.

A few programs in base system needed to be changed due to having their own getline function, most of which aren't anything like getline(3). Hopefully there won't be much fallout in  pkgsrc as a result.

getline(3) is prefered over over functions such as fgetln(3) and fgets(3) because it's standards based and you get a dynamic buffer for really really long lines. However, POSIX did drop the ball on making it a standard from the GNU extension - it should return 0 on EOF and more importantly be called fgetline. Oh well.

I shall be rolling getline(3) support into  dhcpcd later, but I'll have to do a link test in the Makefile to see if we can use it. I'm unsure if I want to have a mini configure for dhcpcd or to keep using just make extensions ....

PIM is the acronym for Personal Information Management: todo lists, email, rss, calendar, contacts, journals, blogs, etc. Recently I have been poking around trying to achieve the “zen” of PIM, where my PIM data is accessible from anywhere, and from any medium – from the internet, from my Windows Mobile 6 powered phone, and from my desktop.

As a KDE-user, naturally I have attempted to use the Kontact PIM suite. KMail and Akregator both work wonders with my data and are no problem, but working with contacts, calendar and to-do lists are a real PITA. The interface for managing the actual PIM storage (which I’m more interested in than the PIM data itself) is completely unintuitive, making me choose from several backend types with no description whatsoever, to work-in-progress Akonadi migration of which the status is quite unknown to me, to random remote/local synchronisation of untitled .ics files. The journal section seems to serve no purpose whatsoever.

Are there any kind souls who have reached their own personal “zen” of PIM management who care to share their setup with me? The criteria is:

1. Hopefully able to use with Kontact
2. Hopefully able to use/synchronise one way or another with my Windows Mobile 6 phone
3. Not using Google services, but a way to store a compatible-with-other-apps file on my personal server would be a definite plus!

Note that I do not necessarily need a feature packed application. For example for a calendar all I want is the ability to say “this happens on this date”, with an option for start/end time. The repeating event feature is also optional but appreciated.

Related posts:

1. Bing.com – another search engine from Microsoft
2. How to Actually Use Your Computer: Part 2
3. Remember The Milk: A Great Online To-do List Service

There'll be a more fulfilling entry in a day or two, after I get my sessions working again. For now, I leave you with a non-Linux related video of a friend...

We need to get the video up to 1000+ views so he will perform again... so please pass this on. For those that can't see it, go here (youtube.com).

Starting today, I’ve had enough with accidentally missing my middle clicks on links and ending up having a (seemingly) random page pop. I should have looked at this a long time ago, but now is a good a time as any. By going into about:config and changing “middlemouse.contentLoadURL” from “true” to “false” I am now rid of this annoyance.

As many people know, I am the layout editor of my school’s “Perspective” magazine. It is a student run organisation and this will be the last issue I design before I hand over my role to the year below (it’s a yearly thing).

I am happy and proud to announce what I believe is the best issue I have ever produced, and you yourself can compare it to the first, the second, and the third issue.

Perspective is made using free and open-source software including The GIMP, Scribus, KDE, Okular, and Vim. However as the industry standard is the proprietary format Adobe InDesign, I am required to convert it to this format at the final stage. However rest assured this is nothing more than copy and pasting – I present to you a magazine made (almost) completely with free software.

This issue is special because you can download this magazine in PDF format. Feel free to read it – it includes a lovely front-page article by me, 3 entires into the art pages at the end, as well as a two-page article about open-source nearing the end featuring pictures of KDE and Elephant’s Dream – the open-source movie by the Blender Foundation. Some kid also wrote an article about the history of web browsers, but I was quite shocked to see that one line said “Google Chrome was released as a beta in September 2008 by Microsoft” – I think they meant for Windows. Nevertheless, my job is to bother about the design, not the standard of articles, and I’m happy to say that this has upped the bar – from what I see at least.

Clickety here to download.

Oh, and for the lazy, here is the thumbnail view of the entire magazine.

Thank you for scrolling through, I hope you’ve enjoyed the magazine over the year, goodbye and good luck to whoever replaces me.

Related posts:

1. Perspective in progress
2. Perspective Nov 2008 Distributed
3. Perspective Magazine Feb 09 Released!

Recently I forced myself to uninstall Amarok 1.4 and try Amarok 2 again. I saw there were some nice updates to the interface coming in the next version so I grabbed the latest version from SVN.

I very quickly started looking for other alternatives, and you'll soon see why. The best I could find was Songbird.

I'll start with a disclaimer that both of these programs are great, and they are free. I am not suggesting, let alone demanding, that anyone change anything in either program to suit me. Kudos and thanks to the devs of both. These two programs are both probably better apps than I could dream of coding. Feel free to respond "Ask for a refund" and "Fix it yourself" anyways if you like. I think it's still useful to give some constructive feedback, and maybe I'll learn something myself about how to make a good GUI along the way.

Next I'll start with my conclusion, so you don't have to read further, because this is admittedly long. Amarok 2's interface is extremely painful, but at least it plays music. Songbird has a wonderful interface, much like Amarok 1.4 had a wonderful interface; if only I could get Songbird to make sound come out of my speakers, I'd be set.

I think it's interesting to compare Songbird and Amarok 2, both being bleeding-edge music players for Linux with a similar philosophy and feature set. So let's compare GUIs. I sized the two windows exactly the same and tried to have them display mostly the same bits of information, so it'd be easy to compare. Click below for larger versions.

Amarok 2:

Songbird:

Playlist

In Songbird the playlist dominates the window by default. This is good because seeing a list of music is what I want. It's the whole point of a music player.

I strongly dislike the "filter pane" style of browsing my music. Thankfully you can turn it off in Songbird. You can also install "cover flow" sorts of eye-candy extensions if that floats your boat. I avoid such things, and Songbird's interface is easy and comfortable by default.

In Amarok by default the playlist is a little sliver of GUI off on the right, and the middle context pane dominates the window. Enough people complained about this that in later versions you can turn off the context view entirely, in which case the playlist will stretch to a reasonable size. Whether the information in it will look good is another story (see below).

Amarok's "Local Collection" browser is an expandable tree. You can customize how things are grouped. This was great in Amarok 1.4. It works similarly here. It's not as lightweight or responsive as in 1.4, but I can't complain. By default it's way on the left, with the playlist way on the right and the context view in between, but in later version of Amarok you can change the order of the panes.

I'll call this a tie even though you have to fight for it in Amarok.

Sorting the playlist

Songbird has a bunch of columns with column headers. To sort things you click the headers. Note that this is how Amarok 1.4 worked. This is how every program in the universe works.

In Amarok you have drop-down menus that you can add and remove with buttons, and you pick sorting criteria from that list, left-to-right in order of priority. This is clumsy. According to the devs' blogs this part of the GUI is a work in progress, which is fine, maybe it'll improve.

But note that the design of Amarok's playlist fundamentally limits the ways you can sort it. There have to be some magic GUI controls floating up top, disconnected from the playlist. You aren't going to get a bunch of column headers that you can click because the playlist isn't just rows and columns. Each song in the playlist can take up more than one row and there are grouping-headers interspersed. This is painful and I imagine it's always going to be painful.

Playlist readability

There are no labels in the Amarok playlist to tell you what information you're looking at in the playlist. I initially customized my playlist to show disc number and track number. Doing so, you get a bunch of numbers. What do the numbers mean? At a glance you can't tell. Am I looking at an Artist or Composer? Play Count, or Score? Does that big empty space mean my song is missing a Genre or missing a Year?

In Songbird the columns have headers.

Playlist length

How many songs can you squeeze into the playlist vertically? This is an important metric for me. I want to be able to find a song quickly without scrolling through a list for a year and a half. Sure I can search, but search doesn't replace my eyes in all circumstances.

In Songbird even with those filter panes above the playlist it fits a few more songs than Amarok. You can turn off the filter panes entirely, in which case you can display tons more songs in Songbird than in Amarok. Songbird wins.

In Amarok, by default the playlist has a bunch of multi-row header stuff mixed into the middle of your playlist to show artists and album names and cover art. You can make the headers not take up so much room (or turn them off entirely), in which case Amarok gets pretty close to Songbird. You'll just do without album or artist names. Unless you can manage to cram them into the playlist in the rows beside the track titles.

Which brings us to our major problem...

Playlist customizability

In Songbird you can right click and add and remove columns. You can drag-and-drop columns to rearrange them. You can drag the edges of the columns to resize them. It's simple and it works. This is how Amarok 1.4 worked too.

Amarok fails hard in comparison. In Amarok to customize the playlist you go into a special dialog. You pick your components from a horizontally-scrolling list of huge icons. Then you arrange them into rows.

You can put two or more items side-by-side in which case they become multiple columns on that row in the playlist. Kind of. To control the width of the columns, you hover over that component in this magical dialog, and a weird circular icon appears. When you click it, a drop-down appears with a microscopic slider at the bottom that looks like it was pulled from KDE2. This is the only way to resize columns in the playlist. Here's a screenshot.

What in the world is this? What are simple drag-and-drop operations in Songbird and every other application ever made, are buried in this cryptic dialog under non-standard controls in Amarok. I've been using KDE and Amarok for a long time and it took me a good couple minutes to even figure out how this thing works.

I think the widths are percentages and have to add up to 100%, I don't even know. The slider is so small that if you drag it one pixel it usually jumps 5-10%, so it's nearly impossible to get anything to look nice. And when you resize the Amaork window later, the columns don't resize sanely; some fields are smashed into each other or overlap as others take up too much space.

Maybe this will all be fixed before the next release; I realize I'm looking at bleeding-edge pre-release software. But this whole idea is so fundamentally broken I don't know how it's going to be salvaged.

I've heard many times that "You can make Amarok 2 look like Amarok 1". No you can't. You can tediously stuff lots of information into the playlist so that it approaches the level of info you could easily and painlessly get in Amarok 1.4. But it will neither look nor act anything like Amarok 1.4. Resizing the playlist will break things. Nothing is labeled. Nothing is easily customizable.

Playlist consistency

Songs in Amarok are grouped into albums by default. If you have a song that doesn't belong to any album, it's displayed completely differently than a song that does. You can alter this in the scary playlist editor dialog mentioned above, under the "Single" tab (as opposed to "Head" and "Body" which control the "grouped" songs). Sound confusing? It is. Needlessly so.

In Songbird songs are displayed the same whether they belong to an album or not, since the play list is just a list of songs. This seems like it should be a no-brainer.

Playlist: overall

Amarok 2's playlist is unique, imaginative, and I'm sure it's a clever bit of code. It's also nearly unusable.

Why can't we have a grid of rows and columns? There's a good reason so many apps use such a control. It's simple and familiar and it works. I'm open to learning something new if it's an improvement. Amarok 2's playlist is not an improvement. Why can't the playlist be a simple list of things to play?

There's nothing about QT4 preventing someone from making a good GUI. Look at ktorrent.

The little things

Say I want to email or IM someone and ask them if they like some artist, whose name happens to be Japanese and difficult to type on my gaijin keyboard. How do you copy and paste the name of an album or artist in Amarok 2? In Amarok 1 you could just click any field in the playlist twice, and it'd let you edit or copy/paste that field inline. Same in Songbird.

In Amarok 2, you have to right click and go into the Edit Song Details dialog, and do it from there, then close the dialog. A tiny step backwards.

How do you change the rating of a song? In Songbird you click the stars in the playlist beside the song you care about. Same in Amarok 1.4.

In Amarok 2, you can display the stars for each song in the playlist, but to change the rating you have to click in the context pane. (So if you dislike and therefore hide the context pane, you're screwed.) Clicking in the playlist does nothing. A tiny step backwards.

All of these tiny steps add up.

Extras

So how well does each player serve as a web browser?

This seems like a ridiculous question, except that both really do try to be a web browser. You can open song lyrics and wikipedia pages and such things right in the music player. I find these features nearly useless. Lyrics are nice when it works (which isn't often, for the music I listen to), but browsing Flickr? Really? Does someone really use this?

Songbird does use its inline browser in a nice way to let you browse and install addons from the Songbird website, and Songbird has a cool feature to let you rip audio files from web pages. Amarok doesn't have these, but I don't hold that against it. I can easily live without any of this stuff.

So in Songbird you have an embedded Mozilla engine. It's hidden behind a tab. You can just avoid opening such a tab and then you don't see it. You can even hide the tab bar itself. Victory.

In Amarok the browser stuff inhabits the middle context pain. The size is limited for this pane, which means information is crammed into the available space, which greatly limits its use. It's also clumsy and difficult to turn components on and off, and I can't figure out how to resize them. The context view itself is either in your face, taking up most of your screen real estate, or it's gone and not easily retrievable.

Note in the screenshot, how in Songbird the lyrics pane is big enough to display all the lyrics, yet small enough not to be annoying. You can also hide the pane (as you can hide every other pane in the GUI) via that tiny button with an arrow under the pane. Amarok's lyrics widget is either too big (if you let it occupy the whole content pane) or too small (if you want to have anything else in the pane with it).

Note that Songbird's lyrics pane is added via an addon. It's a completely optional part of the GUI, which is nice. (Note that Songbird also mangles certain text in the lyrics due to encoding problems, which is a point against it.)

Wasted screen real estate

See that tiny little red icon in the bottom-right of Songbird? That's the Last.fm integration. It's all hidden in a little square of pixels, out of my face, not sucking up screen real estate. This is a common theme in Songbird. Everything is tiny and/or hideable. Tiny is good.

In Amarok everything is huge and round. Even ignoring the content pane, there's white space everywhere. There are buttons strewn all over the interface, like the seven in the lower right. Export Playlist? Does that really need a button? And other buttons appear (and disappear) in awkward positions at the top. "Add Position Marker"? Does this really deserve a prominent button right beside the main play controls?

And yet things I do need buttons for, such as changing the Skip or Repeat options, have no buttons. This is possibly the first player I've ever used that doesn't have a button for Skip and Repeat.

GUI skinning

Songbird is skinnable. So was Amarok 1.4, to a degree. Amarok 2 isn't and I don't know if it ever plans to be. I can live without skins but it's nice to have the option.

Desktop environment integration

As one might imagine, Amarok wins here, if you use KDE, as I do. Global keyboard shortcuts are already set up, it sits in the system tray, and there are nice Plasma applets you can put on your desktop.

Songbird meanwhile does not play nice. First, it has window hints set to hide its border and window title bar, and it tries (and fails) to manage windows itself, giving your window manager the middle finger. I had to force kwin to display the title bar and border just so I could resize certain dialogs that were otherwise broken.

Then, Songbird doesn't sit in the system tray. You can force it down there via alltray, but right-clicking the icon doesn't give you Play/Pause/Next/Back options like in Amarok.

There are no global hotkeys, but you can easily fix this in KDE too because you can set your own global hotkeys to do anything, and Songbird has a commandline interface to let you do what you need. It's still not as graceful as Amarok.

So KDE thankfully rescues Songbird from its own deficiencies, which is nice. Except...

Playing music

Ah, Songbird. Why oh why won't you work? Songbird uses gstreamer. In my years of bouncing between Gnome and KDE and XFCE and others, and using various distros, gstreamer has never worked for me consistently. I can get Songbird to play music, but Flash videos stop producing sound while Songbird is running. This is a known and reported bug, I'm not the only one. While Songbird is playing, other KDE apps randomly produce sound or not depending on the phase of the moon.

Amarok actually plays music, so I'm stuck with it. Unless I go back to Amarok 1.4 which I may still do.

Conclusion

Songbird is pretty good. If I can figure out how to make gstreamer play nice, I'll probably use it.

Otherwise just consider this yet another voice in the wilderness wishing for a Qt4 version of Amarok 1.4. There was nothing wrong with it, from a user's perspective. I'm not the first wishing for this, and won't be the last. If I had a couple years to get good at C++ and a team of programmers to help, I'd probably try it myself.

Why write an 87-page essay about the GUI of a music player? Because Amarok 1.4 was a really good program. I'm a programmer and I appreciate a good program. Songbird has a pretty darned good GUI too. It's painful to see Amarok 2 going in this direction.

Some weeks ago I attended JavaOne (a pretty neat conference, even for non-Java-heads like me) and got in touch with several non-Java languages running on the JVM (nothing really new next to Project Fortress, but I never got into most for real).

Since I wanted to learn some language not resembling any other I already know (even a little), I decided some hours ago to start digging into Clojure, which is a LISP dialect running on the JVM using STM (Software Transactional Memory) and created with concurrency in mind. Check the website for more information.

After some hacking I got a first ‘application’ running. Since recently there’s been some little meme at work regarding echo servers, I decided to write a very basic line-oriented echo server in Clojure.

The result is a server using one thread per connection which just sends back lines to a connected client as-is. Nothing fancy, but might be a useful start for developing basic network applications using Clojure.

Enjoy!

Just a note, I updated my 2008 how to set up a git repository entry. If you're curious, here it is
http://www.mindstab.net/wordpress/archives/288

If you read my initial post about Google Chrome (the OS, not the Brow- wait a minute, is there even a clear distinction anymore?) you would have realised that I didn’t really give opinions on what I felt about it but instead  how I visualised it to be. I believe in designating some mull-over time before making a judgement. (hypocritically speaking, I did not do that when constructing my conspiracy theory when Google Wave came out)

Now is the time to see what exactly is going on.

My feelings in a nutshell

• Would I buy such a product? If it were cheap (100 dollars or so), yes.
• I feel Google is harming open-source.
• Cloud computing is very important to me for accessibility and synchronisation.
• We cannot fight, and should not fight.

The story behind it

The first point is easy to justify and I do believe this is very agreeable. This is an area of the markt people have always looked towards with an expectation of a “trustworthy” brand, and Google has just provided that to them. People will buy for this OS.

To a company, Google is probably executing its marketing strategy in the most effective way possible. They use a product-orientated approach, making the product first then selling it to the market – or so it seems. Google knows two things: 1) They have craploads of data, and 2) They own (pretty much) the biggest mass marketing device in the world. However they do know that even though they “own” this realm, they cannot control it. It’s like a pet – you own but cannot control it.

They way you control it is by feeding it. Such is the nature of open-source development. However Google is able to turn open-source into money by producing a good percentage of the product before open-sourcing it. This allows Google to keep the leash on the project. You developers aren’t building the product side by side – no: you are doing the grunt work that turns a framework into something consumers will love – something with the name Google slapped onto it.

Let’s move onto my third feeling. This is because of a trend I have noticed over time. Computers is no longer about being in full control of your data – it’s about being in full control of your data no matter where you are. Cloud computing sorts this out – it’s no wonder Google’s objective is “to be the hub through which all the world’s information passes through“. Sorry guys, but the fact is that most consumers want this. The only time they won’t is when the company providing it has a bad reputation – but Google? No, Google’s never been evil have they? Not to the average joe they haven’t. It’s the average joe that changes the workflow – it’s the average joe that makes such a way of working part of your daily routine.

You see, Chrome isn’t about making an operating system to do useful stuff – Chrome is all about changing people’s workflow to become web-centric. Instead of moving into the desktop market, what Google is doing is moving consumers into the web market.  Why do you think it’s named Chrome after their browser? It saves on the advertising costs. You advertise the OS, you advertise the browser. Google is pushing ahead HTML 5 specifications to redefine what the web is capable of, and their browser Chrome going to be the biggest, baddest boy in the playground that knows the meaning of the word “compatibility” backwards. Advertise them both at the same time – what you get are people getting the “wow” experience Google can provide with all its toolkits online from the browser, and making it easy as pie to integrate it into how they work. It’s not because Google Docs is simply an application that allows you to edit documents online, it’s because it’s a shared, accessible, compatible, synchronised alternative.

We cannot and should not fight.

Yes. My last point is so awesome it deserves its own special section.

You cannot fight once a market leader has made a choice on a product/system. We saw it with Windows and we may very well see it again. (I assume you have all seen Google Wave?) Instead we have to understand the market. What does the market want? How do we provide for it?

Now, I am a KDE user myself but what I see as major areas for Linux and DEs in general to focus on are:

• Plasmoids (in KDE at least) – this is a stepping stone to integrate new technologies and the web into the desktop workflow
• Provision of private clouds, complying with open-standards – for private, secure and PERSONALISED (imagine giving users the freedom to shape their cloud environment) mobility and synchronisation
• The social desktop
• The semantic desktop

Am I right, am I crazy, have I missed out stuff?

Shower me with your thoughts please.

Related posts:

1. The Google Operating System – Chrome.
2. Beware of Google.
3. How to install Chromium (Google Chome) on Gentoo Linux

All this talk about Google's Chrome OS has made me think a bit.

While in general I agree with both Dion Moult and Christian Weilbach and am in general mistrusting to cloud computing (at least in its currently most popular form), there is something else that bothers me with this hype.

As of late a lot of talk and effort about cloud computing was being done in the direction of making it work on netbooks and similar mobile devices. To me this makes no sense! Having your documents and data online is a great idea if you have no computer of your own or you have to migrate a lot, but only have a stationary system.

But if you can take all your data with yourself on a ultra-portable device with hours of autonomy time (e.g. netbook, smartphone), why would you rather have it online out of your direct reach? It is neither practical, because you need a network connection all the time, nor does it spare you much diskspace (if we completely ignore IP and privacy complications). Yes, SSD's are still small, but frankly, all that music and movies take up a lot more space then your calendars, inbox and documents put together!

There are cases when clouds make sense. But it always depends on what they are used for and how they are implemented. For example, I am very happy with SpiderOak's encrypted and clouded backups and am looking forward to SocialDesktop if its done right (and it seems like it might be). I even have some insane ideas about binding together NEPOMUK with P2P and F2F technology. But I will write about all this some other time.
<!--break-->

I've finally gotten around to some long overdue server maintenance on Kvasir (mindstab.net et all). This is where running a Gentoo server can kind of be fun. I can mostly ignore it for ages, just poking at it when a GLSA (Gentoo security advisement) comes out for software it's running, and then when I feel like sitting down to it, I can upgrade all the software it's running to the latest stable versions.

The difference in comparison to other distros is of course that you are always current with their stable, but their stable can only stay current so long before they have to release a new version or else everything will break on some upgrades. In Gentoo they give you the tools to deal with it and pass the breakage onto you. So I upgrade slowly and cautiously and then fix a few things when config files change or libraries change and more things need to be recompiled, but it works. I installed Gentoo on this server 4 or 5 years ago and look, it's still going and running all new software. I think that's pretty cool.

And to top off the changes, I finally got around to installing a "new" 512MB ram stick in the server as well, now doubling its ram to 1GB. Which is cool, and just in time because the new clamav is eating RAM like candy. I'm actually wondering if something is wrong with it because it's really eating ram...

Ah well, anyways, the server got some well needed love and attention and is feeling much better.

The following configs can be used when you have either Vodafone Mobile Internet or Cosmote Internet on the Go or both 3G USB sticks and you want to connect to the 3G Internet (in Greece) while using Linux. I’ll provide two ways to connect to 3G, the command line way using wvdial and the GUI way using umtsmon.

1) Using wvdial
Create /etc/wvdial.conf:

[Dialer Defaults]
New PPPD = yes
Dial Command = ATDT
Dial Attempts = 1
Modem = /dev/ttyUSB0
Modem Type = Analog Modem
ISDN = 0
Baud = 460800
Username = user
Password = pass
Init1 = ATZ
Init2 = AT&F E1 V1 X1 &D2 &C1 S
[Dialer cosmote]
Phone = *99#
Stupid Mode = 1
Init3 = AT+CGDCONT=1,"IP","internet"
[Dialer vodafone]
Phone = *99#
Stupid Mode = 1
Init3 = AT+CGDCONT=1,"IP","internet"
[Dialer vfPIN]
Init4 = AT+CPIN=1234
[Dialer cmPIN]
Init4 = AT+CPIN=5678

WARNING: You HAVE to change the PINs on the last part of the config

To connect to Cosmote, plug in the usb stick:

# wvdial cmPIN
# wvdial cosmote

To connect to Vodafone, plug in the usb stick:

# wvdial vfPIN
# wvdial vodafone

2) Using umtsmon
Connection->Manage Profiles and create the necessary profiles with settings that look like these:

Username and Password does not really matter. Enter something like User/Pass or Username/Password.

Both versions tested on Debian and Gentoo and they are working just fine.

If someone has the Wind ADSM settings please provide them as a comment so I can complete the post with all three Greek 3G providers.

References: List of AT commands

Read Google’s original blog post about it.

That’s right, my conspiracy theory about Google (orignally posted a good month back) has come true, and it’s going to be out there around late 2010.

Brief summary: Google is making an operating system (Linux-based too) with help from the open-source community that focuses on getting the user online and into a browser as quick as possible. The browser is now the ultimate tool on the system. It is currently mainly meant for stuff like netbooks (note this is a separate project from Google Android) but will apparently also be able to provide a good experience for any desktop setup.

Since it’s too late to grimace at Google during their drawing board sessions, I like to ask myself what would an OS be in a time when many of our activites are web-centric.

Most of the main problems I outlined for Google in my conspiracy theory was how they could convice people to change their workflows. Apparently Google has decided to give them an operating system. This interface can easily be optimised to make it feel natural to shift their workflow completely into what they can do in a browser, some tabs and the new shabang HTML 5 will come with.

I took a look at Moblin, another netbook Linux-based OS – one thing instantly popped through my head: this doesn’t look like any window manager, it looks like a website or single application. Something you might expect similar to MythTV. (If I am wrong please correct me).

The first decision I would make on designing a UI for Google’s purposes is not to have any start menu. Something similar to Apple’s dock with modifications (also with an auto-hide) would be great for optimising screen real estate. I would also integrate what I now see as KDE Plasmoids as part of the entire interface (as in within applications itself too instead of only the desktop shell). I would also ask myself what applications could be and should be replaced by web applications. Such examples are email, document editing, chatting, and social networking. What could not and should not be are graphics and multimedia editors, games, and system management tools. It seems very much now that we can split our activities into 2: if you want to make technology, do it offline. If you want to use technology, do it online.

Personally, I can easily now see how easily I can adapt my workflow to this internet-centric pattern.

What about you? What do you expect from Google’s OS?

Related posts:

1. Chrome in the Clouds: The Google OS
2. Beware of Google.
3. How to install Chromium (Google Chome) on Gentoo Linux

I'm finally getting around to rebuilding my server. I've got an Intel Q6600 CPU, 8gb of ddr3 ram, and five 160gb hard drives for a raid5 array. On top of that, I have two tv tuners if I set up a MythTV server.

At the moment, I haven't decided on a distro for it. I'm by far the. Most familiar with Gentoo. However, it has never been the best choice for servers. So I'm more or less between Ubuntu and Debian as final choices. Suggestions are very much welcome.

The server is mostly going to be used for storage purposes, so I'm fairly sure I'm going to rely on ext4 as the filesystem for my raid array. Again, it's still up in the air. Off the top of my head, I don't entirely remember what other services I wanted to run on it... But it isn't a machine I plan to leave on 24/7 for the moment. I'd really like to build a mini-itx Atom machine to take care of email and other network services for my home.

So more updates will come from this in the near future as I finish assembling the machine itself.

I was originally planning tomorrow to post the new Perspective Magazine for you to ogle at (distribution on Friday!) but I shall delay that and you shall be rewarded for waiting with a PDF you can download of the magazine. Ooh.

Instead tonight I had some free time (because tomorrow I shall be skiving school spending time productively at home) So I decided to look at an NRICH problem. Yes, that’s right. I was so bored I decided to do math. Edit: Apparently I didn’t notice the star rating system so it seems as though I picked an easy one.

Click to see problem: Succession in Randomia

Let’s consider a probability tree:

… OK. The first thing we notice is that it looks prettier. Let’s see the series for B now: 1/4 + 1/16 + 1/64 + … . This is identical to T. We can do the same for L and say we have 2/8 + 2/32 + 2/128 … which is exactly the same as B and T. Plugging it into a/1-r (r being 1/4) we get 1/3 for all three. Hence we can say that in fact it is a completely unbiased way of choosing a successor.

My answer to that question is therefore: “Yeah they have an equal chance”

Well, I wasn’t going to stop here. Why don’t I simulate it? Here is the coding that simulates the situation (the cheapest brute force technique to tackle the problem):

If you don’t know programming, go take a read through the code anyway and see if you can get a grasp of what’s going on

<?php
// The number of times each king as won.
$bingo = 0;
$toto = 0;
$lotto = 0;

// How many times we are going to do the crowning ceremony
$ceremonies = 10000;

// Loop through these instructions to carry out the crowning ceremony.
for ($i = 1; $i < $ceremonies+1; $i++)
{
    $win = FALSE; # Nobody has become king yet.
    $oddeven = 1; # Because the first throw is odd. Assume 1 = odd, 0 = even.
    $firsttoss = TRUE; # We are tossing for the first time.

    // This variable is used later to determine which variable $x or $y is
    // reassigned a value and which keeps the previous toss value.
    $a = 0;

    // Keep on tossing the coins until somebody wins the kingdom.
    while ($win == FALSE)
    {
        // Here we only flip one coin, this will alternate between $x and $y.
        // The one ($x or $y) that isn't assigned a new value will retain the
        // previous toss value so we can find out whether or not we have got
        // two HH or TT in a row.
        // Assume H = 1, T = 2
        if ($a == 0) {
            $x = rand(1,2);
            $a = $a + 1;
        } elseif ($a > 0) {
            $y = rand(1,2);
            $a = 0;
        }

        if ($firsttoss == TRUE) {
            // Nobody can win on the first toss.
            $firsttoss = FALSE;
        } else {
            // If it is HH and ODD...
            if ($x == 1 && $y == 1 && $oddeven == 1) {
                // Lotto will become King.
                $lotto = $lotto + 1;

                // Somebody has won the Kingdom.
                $win = TRUE;
            }

            // If it is TT and ODD...
            if ($x == 2 && $y == 2 && $oddeven == 1) {
                // Lotto will become King.
                $lotto = $lotto + 1;

                // Somebody has won the Kingdom.
                $win = TRUE;
            }

            // If it is HH and EVEN...
            if ($x == 1 && $y == 1 && $oddeven == 0) {
                // Bingo will become King.
                $bingo = $bingo + 1;

                // Somebody has won the Kingdom.
                $win = TRUE;
            }

            // If it is TT and EVEN...
            if ($x == 2 && $y == 2 && $oddeven == 0) {
                // Toto will become King.
                $toto = $toto + 1;

                // Somebody has won the Kingdom.
                $win = TRUE;
            }
        }

        // ... If nobody has won anything ...
        if ($win != TRUE) {
            // We flip again, so Odd->Even or Even->Odd.
            if ($oddeven == 0) {
                $oddeven = 1;
            } else {
                $oddeven = 0;
            }
        }
    }
}

// Tell the computer to print out what results we have.
echo 'Bingo: '. $bingo .'<br />';
echo 'Toto: '. $toto .'<br />';
echo 'Lotto: '. $lotto .'<br />';
?>

I ran it and here is an example result I get. Out of 10,000 ceremonies, this is the number of times each king has won:

Bingo: 3322
Toto: 3340
Lotto: 3338

Pretty close, eh?

No related posts.

I posted recently about being dissatisfied with how my blog was playing with MySQL. I was going to try for some kind of file-based storage, but in the end I decided to go with Tokyo Cabinet, which is a very lightweight key/value store.

This simplifies a lot of code, because a blog is pretty much one list of objects (posts) with a bunch of sub-objects (tags, categories, comments). An SQL DB is made for independent objects that are related to each other via keys. So storing a post means deconstructing it into its parts and stuffing all the parts into their own tables, and fetching a post means fetching all the parts and putting it back together. A bit of a bother.

With a key/value store, a post is a hashmap, and it has sub-lists of comments and tags etc., and you serialize the whole thing and stuff it into the DB; when fetching you un-serialize it and you're good. Clojure being a Lisp, there's already a nice serialization format (s-expressions), so there's hardly any work to be done.

I'll give it a few days and if everything actually works, I'll push the code to github. The code is still pretty nasty and ad-hoc; I did this rewrite in the matter of a couple of hours on a weekend. But parts of it may be useful to someone.

I also plan to do some kind of simplified "How to make a blog in Clojure" tutorial in the near future. The code for this blog is bloated with a lot of functionality that is pretty specific to my site and that most people don't need. A more to-the-point tutorial would probably be more helpful.

I've released cl-pack 0.1.1. Just a few code improvements suggested by Zach. This is just light stuff, some slight tweeks to be more in line with lisp standards and a few small speed up using native functions I didn't know about instead of my own code.

He'd suggested some more radical changes too, the biggest being to turn the two pack/unpack case statements into macro-foo so that I could just go

 
(define-directive #\i ...)
 

all on its own. I gave it a good try and pretty much got it working, but it unfortunately spits out about 125 style warnings, increased the code by about 100 lines or ~ 20% and I'm not sure in the end if it really was any clearer, so I scrapped that for now and am just releasing the basic easy changes and moving onto a few more features.

The planned release will be hopefully something like a few more features for 0.2.0 and then if nothing explodes, release it as 1.0 and cl-pack should be done-ish.

So, I'm now finally natviely IPv6 enabled :)

This led to an  interesting debtate about my internet connection. In a nutshell my  Drayetk Vigor 100 is broken, but apparently by design. The problem is this - for IPv4 goes through my NAT whose public IP has an MTU of 1492 so Path MTU Discovery works fine. However IPv6 does not need NAT and the PPPoE interface does not have a public IPv6 and nor do the internal clients use it even it there was one. So the path MTU is 1500, which is too big for the PPPoE to handle.

The correct solution is to obviously terminate the IPv6 on a PPPoA connection, but only the  Cisco 877 router does this and that's outside of my price range, even on eBay. One solution is to clamp the MSS on the router to 1430 so that all clients work. But this is a hack. The best solution with my hardware is to force the MTU to 1492 for all nodes inside my network, so they share the same MTU as the PPPoE so that Path MTU Discovery works.

You can see this working for yourself on IPv6 now, as this site is IPv6 enabled and sits behind the PPPoE link. You'll need to query IPv4 servers for the address though as I don't yet have a glue record for IPv6, but as I'm changing registrar that should change in a few weeks :)

Because of this, I've made the dhcpcd default to request and use the MTU value if offered by the DHCP server. You'll be seeing this in dhcpcd-5.0.5 (now out). dhcpcd-5.0.6 will feature restoring the MTU correctly between leases.

After a lot of trial and error today, I’ve still only gotten part of the way to my objective: making my new Plantronics USB headset get auto-detected in Gentoo, and make it my “primary” ALSA device. That is, when it’s plugged in, all audio goes to it, and when it’s not, all audio goes to the speakers. Much, much easier said than done.

The first hurdle was coming to the realization that ALSA in fact sees this headset as its own sound card. Once I got that far (it took me a while, with some help on IRC), all I needed in the end was 4 additional lines in my /etc/modprobe.d/alsa file:

alias snd-card-0 snd-hda-intel
alias sound-slot-0 snd-hda-intel
options snd-hda-intel index=1

alias snd-card-1 snd-usb-audio
alias sound-slot-1 snd-usb-audio
options snd-usb-audio index=0

The first 2 lines were already there, and then I added the two lines about snd-card-1 and sound-slot-1. Easy enough. The other two lines (not counting the whitespace) are to tell the system what order they go in. 0 is primary, 1 is second, etc. So by having index=0 for snd-usb-audio, that device is my first card, and the on-board is my second. Easy enough. I kept getting fouled up in testing my various configs by not actually removing the modules; I was just restarting ALSA. Not good. Once I got that config working, I wrote a couple of bash scripts to flip those variables, update the configs, etc. Here's my "on" script:

#!/bin/bash
# Script to swtich around audio devices when headset plugged in
if ( grep -q "options snd-usb-audio index=0" /etc/modprobe.d/alsa )
then
echo "Exiting..."
exit 1
fi

sed -i '/options snd-hda-intel/ s/0/1/' /etc/modprobe.d/alsa
sed -i '/options snd-usb-audio/ s/1/0/' /etc/modprobe.d/alsa
update-modules -f
/etc/init.d/alsasound stop
sleep 0.5
modprobe -r snd-usb-audio
modprobe -r snd-hda-intel
/etc/init.d/alsasound start
sleep 0.5
/etc/init.d/mpd start

It's simple enough, in retrospect, and it works. The "off" script is identical except it does the reverse flip at the beginning. I could have made it one script, but whatever.

The next step was to make this all happen automatically. That's where I'm stuck. I've been tinkering with udev for hours now, and I can't seem to write the rules right. My current rules look like this:

ATTRS{id}=="U0x47f0xc001", ACTION=="add", RUN+="/home/dfego/bin/udev-headset-on.sh", ENV{IS_PLANTRON}="yes"
ENV{IS_PLANTRON}=="yes", ACTION=="remove", RUN+="/home/dfego/bin/udev-headset-off.sh"

I don't even remember now if this particular "add" works, but some of the ones I wrote today did. However, not one, not a single one of my "remove" lines worked. All day. None of them. It's almost depressing. For some God-forsaken reason, I can't make anything trigger on a remove event. So I tried using a single script and having it go on all events, but that fell on its face because udev insists on running it lots and lots of times every time an event happens, even when I built protections into the script that it couldn't run more than once simultaneously. So I don't know what to do. For now, I'm just going to be happy with what I've done and use the script. It's not like I plug my headset in and out all that often (except, of course, today). But I feel sad and defeated, and very much like I wasted a ton of time doing something that doesn't work for some reason beyond my comprehension. It should work... it just doesn't.

Maybe I'll try again one of these days. Maybe not.

O Firefly (antigo mt-daapd) é uma implementação DAAP (Digital Audio Access Protocol) que é o protocolo utilizado pelo iTunes para compartilhar as bibliotecas de audio na rede. É compativel com iTunes no Windows e no Mac OS X assim como players compativeis com DAAP como o Amarok. Portanto é uma boa idéia manter um repositório de mídia centralizado em um servidor na rede.

A homepage do projeto é http://www.fireflymediaserver.org/. Além de servidor DAAP o Firefly provê mais alguns goodies. Vale a pena checar.

Instalação:

A instalação no FreeBSD é bastante simples. Considerando que voce tenha o ports (e saiba usa-lo) apenas instale-o a partir de /usr/ports/audio/firefly. Não há qualquer problema em manter os padrões sugeridos para as dependencias.

Configuração:

Adicione “firefly_enable=”YES” e edite o arquivo de configuração /usr/local/etc/mt-daapd.conf. Lembre-se de criar as pastas usadas para a database e certificar-se de que pertencem ao usuário daapd. Inicie o serviço via /usr/local/etc/rc.d/mt-daapd start. As configurações e atualizações manuais da biblioteca de midia podem ser manipulados pela interface web http://<server>:3689.

O firefly se anuncia na rede utilizando o Bonjour da Apple e pode conflitar com o Avahi.

Se voce completou os passos descritos acima com sucesso, poderá visualizar sua biblioteca de audio nos computadores de sua rede local através do iTunes ou Amarok.

Hello everybody, I’m back from my 5 day jungle trek and I’m just catching up on what I’ve missed throughout the week. I was initially going to award you all with a post about the trek itself, but it turns out Jonathan Williamson from Montage Studio (the very same who does the Blender screencasts and gave some good tips for ThoughtScore) has got himself a Blender build for Windows 7 and has recorded a short screencast demo-ing the development.

I am truly amazed with what has been going on and I will definitely throw myself back into Blender this holiday and its stuff like this that really shows what open-source is capable of. Blender is one serious threat to the huge commercial monopoly in the 3D industry. Here is a short list of the features he describes:

• New design/look
• Panel splitting/deleting/management
• Not limited to one window only
• Massive reorganisation of features that make it more intuitive
• Real-time playback animation while editing
• Real-time playback animation while rendering
• Every single value in Blender can now be animated
• Support for macro options
• New transform panel
• Search option for features

Without further ado:

Clicky here to watch the video.

Related posts:

1. The Blender Model Repository and BlenderNation: open-source merger?
2. Blender 3D: Architecture, Buildings and Scenery – Review
3. Blender Suzanne Awards announced.

Real languages don't have reserved words

proc if {cond expres} {
 puts "cond is $cond"
 puts "expr is $expres"
}

set a 1
if {$a == 1} {
 puts hello
}

proc open {} {
 set ::alarm_socket [open_socket $::options(-alarm_host)]

 foreach host $::options(-hosts) {
  verbose "open $host"
  set ::sockets($host) [open_socket $host]
  # set up an event handler for when data is readable on this socket:
  fileevent $::sockets($host) readable [list process $host]
  # initialise socket timeout with open time
  set ::socket_t($host) [clock seconds]
 }
}

$ ./test.tcl
invalid command name "::tcl::tm::UnknownHandler"
while executing
"::tcl::tm::UnknownHandler ::tclPkgUnknown msgcat 1.4"
("package unknown" script)
invoked from within
"package require msgcat 1.4"
("uplevel" body line 2)
invoked from within
"uplevel \#0 {
package require msgcat 1.4
if { $::tcl_platform(platform) eq {windows} } {
if { [catch { package require registry 1.1 }] } {
..."
(file "/usr/lib/tcl8.5/clock.tcl" line 23)
invoked from within
"source -encoding utf-8 [file join $TclLibDir clock.tcl]"
(procedure "::tcl::clock::format" line 3)
invoked from within
"clock format [clock seconds]"
(procedure "alarm_timeouts" line 3)
invoked from within

   puts "$::argv0: [clock format [clock seconds]]"

Real languages don't have types:

set a world
puts "Hello $a" ;# prints 'Hello world'
set a 1
incr $a
puts "$a + 2" ;# prints '2 + 2'

set a pu
set b ts
$a$b "Hello World"

Hello World

Real languages don't have comments

# if ($sometest) {
 $somecode
#}

if (0)

Conclusion